ASQ CQA – 5. Quality Tools and Techniques Part 3

5. 5A4 Statistical Process Control (SPC) Charts

Is control chart. Control chart is basically the measurement of an item over a period of time. So you keep on measuring and then you keep on recording based on the time. And in addition to that, this chart has control limits as well. These control limits are calculated using some statistical formulas. Now, why we need control chart? We need control Chart to display and manage where variation in the process over time. Let’s say if you’re doing machining of an item and then slowly over time this item dimension gets changed. This is what you want to see.

That because of tools we are and here the item dimension is getting changed. So you need to catch that variation much earlier rather than making defective pieces and then catching that. In addition to that, why do we need Control Chart or SPC? We need it to identify when a process changes. You need to catch that change quicker rather than waiting for a lot of things being produced and then getting rejected. Another important thing in statistical process control is the cause of variation. The cause of variation could be the common cause or special cause. Here we want to identify whether this change is because of common cause or special cause.

We will be talking about these two causes later on in this video. And then as I earlier told, the purpose of Control Charts or SPC is to tell operator when to stop and when to keep on going. If things are changing at what time someone should stop and relook at things, relook at the setting, make the changes in the setting and then restart the process or let it go because this variation is general variation or this variation is because of common causes. To understand control charts or statistical process control, let’s take a simple example. So in this example, let’s assume that I have to reach office at 630 in the morning. So I start and do some traveling and I reach office at 630. And as you know, because of number of things such as traffic signal light or weather or other things, the time it takes me to travel changes.

So there is always a variation. So whether I make a piece or whether I travel, there is always a variation. And because of this variation I might not be able to reach office exactly at 630 every day. Some days I will reach slightly earlier, on some other days I will reach slightly late. And this is what we are trying to understand the variation in the time. So here if you see I have a simple control chart here. In this control chart the average is 630 in the morning. That’s my mean time. And then I have a plot which is going up and down. This is my actual reaching time on day number one to day number 18. And as you would see that there is some variation. But generally looking at this I can say that okay.

I cannot reach office at 630 exactly, but I reach somewhere between 610 and 650. If I reach earlier, I am reaching around 610. And if I’m delayed, I reach around 650. This is what I can see from this control chart. And this variation in reaching time could be because of two types of causes. Common cause and special cause. We will understand the difference between these on the next slide and that will basically help us in deciding whether I need to take some action or I need to just let it go. So here on this slide I have two types of causes of variation. The common causes which are on the left and special causes which are on the right. Let’s talk about common causes. Common causes are many causes and they have minimum impact. In case of me traveling from home to office, common cause could be traffic. Someday traffic is high. Someday traffic is low. Someday I get two red lights. Someday I get three red lights. So these are common causes. There are so many common causes. You really cannot do much about that. On the other hand, there are special causes.

Special causes are few causes which have significant impact. Let’s say one day there was a heavy snowfall. The roads were slippery, vehicles were slipping. So on that day, let’s say I am not able to reach even at 650. So on that day, let’s say that I reach office at 720. This was much more delayed compared to the normal fluctuation which I had. My normal fluctuation was between 610 to 650. Because of very special cause, heavy snowfall, slippery roads, I reach office at 720 or 730. So that is a special cause. You can do something about special causes but it’s really not economical to work on common causes. So these two causes, common causes and special causes are also known by a few other names. For example, when it comes to common cause, common cause is known as random cause. Because these are random. This is also called as chance cause.

Or this is also called as non assignable cause because you cannot assign that particular problem to something very specific thing. So this is the reason. This is called as non assignable cause as well. On the other hand, special cause is called as the signal which tells that when to act. This is also called as systematic or assignable cause. Here you can assign the cause to a specific reason. That day when I reached 720 to my office, there was a special cause. I could assign my delay to a specific cause that just because of heavy snowfall, slippery roads, I got delayed.

So these are two causes which will help you in deciding when to act, when to not act. Common causes are something which are natural. So you really cannot act on that. But once a special cause comes into picture into your process, that is the time you should act. Now, there are a number of control charts and we will not go into those details. Here we have the list of control charts, control charts for variables and attributes. And when I say variable, that means measurement, where I take measurement, measurement of length, measurement of time, measurement of volume.

So when I take measurements, in that case, I can use these three charts, the top three IMR or XMR, these are called as individual and moving range chart. Previously, the example which I showed you, which was the control chart for me reaching office, this was one part of IMR chart, individual value and moving range. These are basically a pair of charts. So here you basically use two charts. One chart for individual value and one chart for moving range value. The chart which I showed you earlier was individual chart that was charged for the time I reached my office. That did not include the second part which was the moving range chart. Similarly, I have X bar R chart, x bar is average and R is the range, x bar S chart, x bar is the average and S is the standard deviation.

So these are three types of charts, or rather I should say pair of charts which you can use when you are measuring something as against measuring. When you’re counting something. Then you can use these four charts at the bottom which are NP chart. NP chart is number of defectives p chart which is for proportion defectives, c chart which is for number of defects and U chart for number of defects per unit. So, if you look at these four charts, the top two charts in that are related to number of defectives. Here you are counting number of defectives. How many things are wrong? The bottom two out of these count charts are based on number of defects.

One item which is defective could have number of defects. So, depending on the situation, we can use different charts. We are not going into actual details of these charts, we are not going into the calculations of these charts because this is beyond the scope of CQA body of knowledge. These things are discussed in much more detail when you are looking for CQE, which is Certified Quality Engineer or Six Sigma Green Belt or Black Belt courses. So there, this particular topic is discussed in much more detail.

Now, quickly, let’s look at those charts which we saw earlier, where and how we use. So, as we saw earlier, depending on the data type, whether the data is a measurement or a count, you have different charts. So if the data is measurement, so we follow this path and if data is count, we follow this path. In case of measurement, we can use these charts IMR or also called as XMR, x bar R and X bar S. Which one we use? That depends on subgroup size. If subgroup size is one, you use IMR chart. If subgroup size is between two and nine, you use X bar R chart. If subgroup size is greater than nine then you use X bar S chart. And what is subgroup size? Subgroup size is number of items you pick at a time. So if you have a production line, you pick five item every hour and check the dimension of that. That number five is your subgroup size. So if your subgroup size is five, then basically you will end up here which is X bar chart. If you are picking 20 items every hour, taking measurement of that, plotting a control chart for that. So for 20 items which is greater than nine, you will be using X bar S chart.

So this was for the measurement part. So if you have count so in counts you need to see whether you are counting defects or you are counting number of defectives. So if you are counting defects, then you follow this path. On this path depending on the sample size, whether your sample size is constant or variable, whether you pick 200 items every time, or maybe sometimes you pick 200 items, sometimes you pick 400 items. In that case your sample size is variable. So if sample size is variable you use U chart. If sample size is constant then you use C chart. This was for number of defects. If we come to number of defectives in that as well you need to see whether your sample size is constant or variable. If sample size is constant then you use NP chart. If your sample size is variable then you use P chart. I understand that I have not gone into much more detail, but probably you don’t need to go into that much detail for your CQA exam. This is all you need for your CQA exam.

Now let’s quickly look at how to interpret these charts. So we make these charts. So what I have here is X bar chart. So previously when we said that when we are measuring something we have X bar R chart, x bar S chart or IMR chart. So out of that I’m just using X bar chart. So here what I am plotting is the average. Now how do we interpret these results of these charts? So one thing is that you have upper control limit and lower control limit. So in the chart which I was using for reaching my office, my upper control limit was 650, my lower control limit was 610. That means in most of the cases under normal circumstances I will reach office at any time between 610 and 650. So those were my upper control limit and lower control limits. Now, one way to look at these control charts is to see if there’s any point which is going above or below the upper control limit or lower control limit. So in our example, when I reached office at 710, this was above my upper control limit.

So that means something has happened. So there was a special cause, there was an assignable cause and something needed to be done. So this was one rule for looking at the control chart, for interpreting the control chart, that if any point goes above the upper control limit or below the lower control limit, then you can say that something is wrong. In addition to that, there are a few other rules as well. So what I have here is Nelson’s rule. So these are the rules which basically tell that when to act. So when you look at the control chart, so you will be looking at a number of things. So the first rule was that one point more than three standard deviation from the center line, which is basically means above the upper control limit or below the lower control limit. So here I see there is a point which is above the upper control limit. There is one point below the lower control limit. So this is one thing. The second rule is that nine or more points on the same side of the center line. So here you see that nine consecutive points are on the same side of the central line.

The third rule is where your points keep on rising. So here we have six or more in a row in the same direction, either increasing or decreasing. That means something is wrong. There’s something happening here that all these points are increasing continuously or decreasing. So there should be some action required at that time. Rule number four is alternating 14 points, alternating up, down, up, down, up, down. So this means something is happening, something is being tampered or someone is basically controlling. If things go up, then this operator is making something so that the next reading goes down and so on. So this was rule number four. Rule number five is there are two points out of three which are more than two standard division from the center line. I’m not going into too much of details in regards to the control chart, but when I say upper control limit and lower control limit, this basically can be divided into three layers.

So one sigma, one standard division, two standard division and three standard division. We will be talking about standard division as we go further and learn about some basic statistics. But then for CQA exam you just need to know some basic things. But right now let’s say this control chart which has upper control limit, lower control limit, there are two more lines, one sigma, two sigma and three sigma. And in the bottom also one sigma, two sigma and three sigma. So if there are two points out of three which are in two sigma range between two and three sigma, which is here, then also we say that something is wrong, something is happening. There is an assignable cause. Similarly the 6th rule is four out of five are more than one standard deviation away from the center line. The 7th rule is that everything is so close to the center line that everything is in plus minus one sigma.

That means something really good has happened. Everything is so controlled, everything is so narrow. Then that means you might have to investigate and check whether there is something which has improved which you want to make a norm for the process. And the 8th rule is eight points in a row, more than one standard division. So you don’t have anything in one standard division. Everything is outside all the eight points. Consecutive points are outside one standard division. So these are eight rules. Some books have different numbers. So instead of nine, you might see some places this as eight. Instead of six, you might see something as seven. Don’t worry about these numbers. Different rules have slightly different numbers, but you need to understand the intent of these. So this is.

6. 5A5 Check Sheets

When we were talking about the Parato chart, in Parato chart, we collected some data that are related to bottles of 300 501,000 milliliter bottles. And we were looking at things like scratch and lose cap and label, et cetera. That was check sheet. So check sheet is a data collection form. And that’s what we were doing there. We were collecting data, data related to the problems which we have in that water bottling plant. So here we have form where we manually tally and record the number of observations or the occurrence of certain events during a specified period. That’s what we were doing for one day duration. We were looking at the production line. And every time I get a problem, I put a telemark. And then why we need this tool? We need this tool to collect data and to display the data. So as an auditor also you might want to use this tool when you’re doing audit. So when you go look at an operator doing something and you want to see like piece one being made, piece two being made, whatever data collection you want to use, you can use this tool there. And this particular tool not only collects the data, but displays the data also.

So you can visually see something, you can make some sense out of that data and solve some problem. Coming to the example which we saw earlier. So this was the example. So here, what you have is a table which has the capacity 300 501,000 milliliter bottles. And here I have the defects which could happen. So what happens is, let’s say I stand there and inspect all these bottles. So when I pick a bottle, that’s fine, I keep it on the side. Second bottle is good, that’s fine. Let’s say 10th bottle, I find a scratch. So what I will do is this scratch was on 300 milliliter bottle.

So let me put a tally here. So I put a tally here. Then maybe after some time I get another bottle which has a problem related to volume. So I put a tally here. Anything related to leakage, I put a tally here. Second time a bottle came with a scratch, I put a second tally here. This means there are two problems related to scratch when it comes to thousand milliliter bottles, looking at the loose cap. So what I do is I put first tally, one problem, second problem, third problem, fourth problem. And then when I get the fifth problem, what I do is I put a cross mark here. So here in this tally, we make groups of five. So this is how we keep on recording this information. And then at the end I can take total of these. So this is how I use my check sheet. Now, there are other type of check sheet as well, where you have something visual. So let’s say I’m inspecting refrigerators. So my company is making refrigerators and these are shipped to the dealers and then dealer is reporting that there are some dents on the refrigerator body. Now what I could do is I can make a sketch of the refrigerator.

So here is my fridge, the bottom portion. Then let’s say two doors here and this is the body of my fridge. So this is something which I can use as a template. So what will happen is next time I see a dent or the scratch or the body mark on the fridge, I put across here. So first time I saw a problem near the freezer, I put across here second fridge or maybe with second defective fridge I find a dancer or something. I put cross at the body sometimes I had here. But then if I keep on recording this over time, then probably I can make a good judgment. You know what, even though there are a lot of dents happening on this refrigerator, but most of these dents are happening in this area, that basically helps me in identifying the problem as well. It not only helps you in recording the problem, sometimes this can help you in identifying the exact location or doing some extra analysis in addition to recording.

One common mistake people make is the difference between the check sheet and the checked list. What we are talking here is check sheet. Check sheet is for data collection. Here we are collecting data, putting a tally mark 12345, or maybe putting a cross here on the body of the fridge. So here we are collecting data as against check sheet. What checklist does is checklist basically reminds you to do something. So checklist is basically a mistake proofing tool or a reminder tool. So when you are doing audit, you use checklist for all the checkpoints which you need to go through. So you will have a checklist which will tell the check this, check that, check this, check that. That’s checklist, check sheet is your recording tool where you record information.